Although the study of the comprehension of prose has made considerable progress, the most promising theoretical models of comprehension, computer simulation models, have not been fully exploited due to lack of a method for comparing the models to experimental data. The purpose of the proposed project is to test an already-developed simulation model of the comprehension of prose. In some preliminary work, a new method based on multiple regression was used to fit the simulation to some experimental data on reading times in simple passages. The results of the preliminary work suggest that the model is capable of accounting in detail for the effects of changes in processing load due to passage organization on the reading times for individual sentences. The proposed work, appropriate for NIMH Small Grant funding, is to collect additional experimental data on reading times that will allow certain assumptions in the simulation to be tested, and the method of fitting the simulation to data to be tried out more completely.